add catboost and tabnet for tabular regression tasks

alpha_automl/resource/primitives_hierarchy.json

@@ -69,7 +69,8 @@
         "sklearn.linear_model.RidgeCV",
         "sklearn.linear_model.TheilSenRegressor",
         "xgboost.XGBRegressor",
-        "lightgbm.LGBMRegressor"
+        "lightgbm.LGBMRegressor",
+	"catboost.CatBoostRegressor"
     ],
     "TEXT_ENCODER": [
         "sklearn.feature_extraction.text.CountVectorizer",

alpha_automl/wrapper_primitives/tabnet.py

@@ -0,0 +1,36 @@
+import torch
+import numpy as np
+import pandas as pd
+from alpha_automl.base_primitive import BasePrimitive
+from alpha_automl._optional_dependency import check_optional_dependency
+
+ml_task = 'regression'
+check_optional_dependency('pytorch_tabnet', ml_task)
+from pytorch_tabnet.tab_model import TabNetRegressor
+
+class PytorchTabNetRegressor(BasePrimitive):
+    """
+    This is a pyTorch implementation of Tabnet
+    (Arik, S. O., & Pfister, T. (2019). TabNet: Attentive Interpretable Tabular Learning. arXiv preprint arXiv:1908.07442.)
+    https://arxiv.org/pdf/1908.07442.pdf.
+    Please note that some different choices have been made overtime to improve the library
+    which can differ from the orginal paper.
+    """
+
+    def __init__(self):
+        self.device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
+        self.model = TabNetRegressor(n_d=32, verbose=False)
+
+    def fit(self, X, y):
+        if isinstance(X, pd.DataFrame):
+            self.model.fit(X.values, y.values)
+        else:
+            self.model.fit(np.asarray(X), np.asarray(y))
+        return self
+
+    def predict(self, X):
+        # Load fasttext model
+        if isinstance(X, pd.DataFrame):
+            return self.model.predict(X.values)
+        else:
+            return self.model.predict(np.asarray(X))

extra_requirements.txt

@@ -6,4 +6,6 @@ mxnet==1.9.1:               timeseries
 pmdarima==2.0.3:            timeseries
 fasttext-wheel:             nlp
 transformers:               nlp, image
-scikit-image:               image
+scikit-image:               image
+catboost:                   regression
+pytorch_tabnet:             regression